full p
aper
High-Mobility Naphthalene Diimide and
Selenophene-Vinylene-Selenophene-Based Conjugated Polymer:
n-Channel Organic Field-Effect Transistors
and Structure–Property Relationship
1. Introduction
Over the past two decades, the field of solution-processable organic field-effect transistors (OFETs) has experienced steady progress as a result of the tuning of the electrical properties by chemical synthesis, the control of the microstruc-ture of organic semiconducting films by optimizing deposition processes, and the careful selection of gate dielectrics and charge injection electrode materials with suitable interfacial properties.[1–7] Such
progress coupled with the high compat-ibility of solution-processable organic semiconductors with plastic or metal foil substrates makes them ideal candidates for cost-effective, mechanically flexible electronic devices for various applications, such as printed radio frequency identifi-cation tags for item-level tagging, drivers for flexible displays, wearable electronics, distributed sensors, and integrated, non-volatile memory devices.[8–20] Recently,
solution-processed p-type OFETs with
Min Jae Sung, Alessandro Luzio, Won-Tae Park, Ran Kim, Eliot Gann,
Francesco Maddalena, Giuseppina Pace, Yong Xu, Dario Natali, Carlo de Falco,
Long Dang, Christopher R. McNeill, Mario Caironi,* Yong-Young Noh,* and Yun-Hi Kim*
Interdependence of chemical structure, thin-film morphology, and transportproperties is a key, yet often elusive aspect characterizing the design and development of high-mobility, solution-processed polymers for large-area and flexible electronics applications. There is a specific need to achieve >1 cm2 V−1 s−1 field-effect mobilities (μ) at low processing temperatures in combination with environmental stability, especially in the case of electron-transporting polymers, which are still lagging behind hole transporting materials. Here, the synthesis of a naphthalene-diimide based donor–acceptor copolymer characterized by a selenophene vinylene selenophene donor moiety is reported. Optimized field-effect transistors show maximum μ of 2.4 cm2 V−1 s−1 and promising ambient stability. A very marked film structural evolution is revealed with increasing annealing temperature, with evidence of a remarkable 3D crystallinity above 180 °C. Conversely, transport properties are found to be substantially optimized at 150 °C, with limited gain at higher temperature. This discrepancy is rationalized by the presence of a surface-segregated prevalently edge-on packed polymer phase, dominating the device accumulated channel. This study therefore serves the purpose of presenting a promising, high-electron-mobility copolymer that is processable at relatively low temperatures, and of clearly highlighting the necessity of specifically investigating channel morphology in assessing the structure–property nexus in semiconducting polymer thin films.
DOI: 10.1002/adfm.201601144
M. J. Sung
School of Materials Science & Engineering and ERI Gyeongsang National University
501 Jinju Daero, Jinju 660-701, South Korea
Dr. A. Luzio, Dr. F. Maddalena, Dr. G. Pace, Prof. D. Natali, Dr. M. Caironi
Center for Nano Science and Technology @PoliMi Istituto Italiano di Tecnologia
Via Pascoli 70/3, 20133 Milan, Italy E-mail: [email protected]
W.-T. Park, Prof. Y. Xu, L. Dang, Prof. Y.-Y. Noh Department of Energy and Materials Engineering Dongguk University
26, Pil-dong, 3-ga, Jung-gu, Seoul 100-715, South Korea E-mail: [email protected]
Dr. R. Kim, Prof. Y.-H. Kim
Department of Chemistry and RIGET Gyeongsang National University
501 Jinju Daero, Jinju 660-701, South Korea E-mail: [email protected]
Dr. E. Gann, Prof. C. R. McNeill
Department of Materials Science and Engineering Monash University
Wellington Road, Clayton, Victoria 3800, Australia Prof. D. Natali
Dipartimento di Elettronica Informazione e Bioingegneria Politecnico di Milano
Piazza L. da Vinci 32, 20133 Milan, Italy Prof. C. De Falco
MOX Modeling and Scientific Computing Dipartimento di Matematica
Politecnico di Milano
Piazza L. da Vinci 32, 20133 Milan, Italy
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. The copyright line of this paper was changed 24 October 2016 after initial publication.
full p
aper
outstanding charge carrier mobilities have been demonstrated, with the best small molecule and polymer OFETs showing mobilities exceeding 10 cm2 V−1 s−1.[21–25] Such remarkable
p-channel mobilities, obtained in devices with fairly promising shelf-life and operational stabilities, have been demonstrated with conjugated polymers based on new building units, such as diketopyrrolopyrrole,[24,26,27] isoindigo,[28,29] and
indaceno-dithiophene.[30] Similar efforts have been devoted to improving
n-channel polymer semiconductors, with the performances of n-type devices still lagging behind their p-channel counter-parts.[31–36] The main obstacle that has traditionally prevented
the improvement of electron transporting polymers is the pres-ence of trap sites forming at around −3.8 to −4.0 eV, mainly associated with hydrogenated oxygen, which complicates the design of conjugated polymers because of the requirement for a lowest unoccupied molecular orbital (LUMO) energy level that lays below this range.[37] In the recent past, different
solutions have been proposed, among which a remarkably well performing solution-processable n-type polymer reported by Yan et al., based on a naphthalene diimide (NDI) acceptor moiety (referred to as PNDI2OD-T2) and easily achieving field-effect mobilities in the 0.1 to 1 cm2 V−1 s−1 range, represented
a breakthrough in the development of n-channel organic poly-meric semiconductors.[38] The two strong electron-withdrawing
imide groups per naphthalene unit can effectively pull down the LUMO energetic level, leading to improved air stability. The electron-deficient NDI center implements a bicyclic pla-narization within the conjugated structure, intended to provide a strong π–π interaction and hence efficient interchain trans-port of intrachain localized carriers throughout the film.[39]
Furthermore, the ease of functionalization at the N-position of the imide ring with different substituents enables control of the physical properties such as solubility, crystallization, air stability, and self-assembly.[40,41] More recently, improved
elec-tron mobilities could be achieved by acting on the donor unit in a NDI-based copolymer. In particular, an electron mobility of ≈1.5 cm2 V−1 s−1 was achieved by copolymerization of a long
alkyl substituted-NDI unit with a bithiophene end-capped thie-nylene-vinylene-thienylene unit (PNDI-TVT).[42]
Alternative opportunities for the basic structure of the donor moiety stem from synthetic efforts that have in the past been directed at replacing thiophene units with electron-rich sele-nophenes, with the purpose of enhancing the intermolecular interactions and thus interchain charge transfer with respect
to thiophene analogues.[43–46] Such a path has led various
research groups to demonstrate high-mobility p-channel conju-gated polymers with improved ambient stability.[44,47–51] Despite
the strong interest in the incorporation of selenophenes in p-channel conjugated polymers, their use in n-channel and/or ambipolar polymers has been quite limited so far.[52]
While the chemical structure of the polymer semiconductor is of fundamental importance in determining the optoelec-tronic properties of a material, it is also very well known that in the solid state the film morphology and the packing motif of the conjugated segments critically size the charge transport proper-ties.[1,2] For example, control of the alignment of the polymer
backbones in a thin film can strongly enhance the charge car-rier mobility.[53–57] In the case of the model n-type polymer
PNDI2OD-T2, an electron mobility exceeding 6 cm2 V−1 s−1 at
high fields was achieved with a fast bar-coating process, starting from solutions with a high degree of preaggregation, leading to aligned fibrillar domains over large areas[3] as a result of flow
induced alignment, achievable with an off-center spin-coating process as well.[58,59]
Such results indicate the beneficial interplay of chemical structure and solid-state morphology in improving charge transport properties, a common aspect of solution-processed high-mobility polymers typically indicated as the structure– property nexus. This relationship has not been fully resolved yet, and while the bulk film microstructure is most typically correlated with transport properties, elusive and sometime con-tradictory aspects have recently induced researchers to inves-tigate more closely the so-called functional morphology, i.e., the morphology of the only volume of the film that is involved in charge transport within the nanometer-thick accumulated channel of a FET.[60–62]
In this contribution, we fulfill the double aim of reporting the synthesis of a new and high-electron-mobility NDI-based copoly mer, incorporating a selenophene-vinyl-selenophene donor moiety, and of evidencing a clear deviation between channel and bulk film morphology. First, we report on the synthesis and full characterization in an OFET device con-figuration of the solution-processed polymer poly[(E)-2,7- bis(2-decyltetradecyl)-4-methyl-9-(5-(2-(5-methylselenophen-2-yl)vinyl)selenophen-2-yl)benzo[lmn][3,8] phenanthroline-1,3,6,8(2H,7H)-tetraone] (PNDI-SVS, Scheme 1). OFET devices based on this new copolymer are characterized by high elec-tron mobility, exceeding 2.4 cm2 V−1 s−1, with a promising
Scheme 1. Synthesis of poly[(E)-2,7-bis(2-decyltetradecyl)-4-methyl-9-(5-(2-(5-methylselenophen-2-yl)vinyl)selenophen-2-yl)benzo[lmn][3,8]
full p
aper
ambient-air operational stability. We then thoroughly investi-gate the microstructure of PNDI-SVS thin films as a function of annealing temperature by means of grazing incidence wide angle X-ray scattering (GIWAXS), showing a marked evolu-tion toward a more crystalline phase with temperatures up to 180 °C, where a remarkable 3D crystal structure arises. Inter-estingly, very good electron mobility can be achieved already at low processing temperatures (150 °C), highly favoring appli-cation of the copolymer in plastic-based flexible circuits. This apparently contradictory result is rationalized through a deep analysis of the micro and electronic structure within the few nanometers-thick accumulated channel at the semiconductor– dielectric interface thanks to near edge X-ray absorption fine structure (NEXAFS) spectroscopy, extraction of the effective density of states, and charge modulation micro-spectroscopy.
2. Results
2.1. Synthesis and Density Functional Theory Calculation
PNDI-SVS was synthesized by a Stille coupling reaction using Pd2(dba)3 and P(o-Tol)3, as shown in Scheme 1. The polymer
was extracted with chloroform, collected by precipitation in methanol, and purified by successive Soxhlet extractions using methanol, acetone, toluene, and chloroform to remove biprod-ucts and oligomers. The chemical structure of the obtained PNDI-SVS was confirmed by 1H NMR (Figure S1, Supporting
Information). The PNDI-SVS is soluble in common organic solvents, notably chloroform and chlorobenzene. The molecular weight of the polymer was determined by gel permeation chro-matography (GPC) with polystyrene standards in chloroform (see Figure S2, Supporting Information). An average molecular weight (Mn) of 80.0 kg mol−1 and a weight average molecular
weight (Mw) of 168.0 kg mol−1, with a polydispersity index of
2.1, were found. Such a high molecular weight was achieved owing to the good solubility in common solvents.
The optical absorption of a PNDI-SVS chloroform solu-tion and thin film deposited from spin coating is shown in Figure S3a of the Supporting Information. PNDI-SVS absorp-tion spectra are characterized by absorpabsorp-tion maxima at 403 and 711 nm in solution, and at 416 and 743 nm in the solid state. Analogously with previous reports on PNDI2OD-T2, the two distinctive peaks observed can be attributed to a π–π* transi-tion and an intramolecular transitransi-tion with charge transfer character between the donating unit and the electron-acceptor unit, respectively.[63–67] With respect to the analogue
NDI copolymer with the thienylene moiety instead of the sele-nophene, PNDI-SVS UV–vis absorption maxima are 30 nm red shifted owing to the introduction of a slightly stronger electron donor unit. Correspondingly, the optical band gap
of PNDI-SVS, estimated at the onset of absorption maxima, is 1.31 eV, slightly narrower than the bandgap of PNDI-TVT. This is presumably owing to the extended π-conjugation achiev-able through the introduction of the electron-rich SVS moiety. As shown in Figure S3b of the Supporting Information and Table 1, from cyclic voltammetry measurements, the estimated LUMO energy of PNDI-SVS is −3.98 eV. The LUMO energy of PNDI-SVS is slightly lower than that of the well-known NDI polymer PNDI2OD-T2 (−3.91 eV), a variation in the direction of improved air stability.[63] The thermal properties of the polymer
were studied by carrying out thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). PNDI-SVS showed good thermal stability, with weight losses stating at 387 °C in the TGA measurements (Figure S4, Supporting Information). PNDI-SVS showed no glass transition in the temperature range of 50—250 °C in the DSC thermograms (Figure S4, Supporting Information).
Density functional theory (DFT) calculations were also car-ried out at the B3LYP 6–31G** level. As expected, the coeffi-cients of the HOMO orbital are located on the SVS unit while that of the LUMO orbital is mainly positioned on the NDI unit, in good agreement with other NDI derivatives (Figure S5, Sup-porting Information).[31,67] The dihedral angles between the
NDI and SVS units are 47.70° and 43.78° (Figure S5b, Sup-porting Information), which are larger than those reported for NDI and TVT, and NDI and bithiophene units, being 33.44° and 32.82°, and 32.88° and 40.61°, respectively.[66,68,69] The
dif-ference between the dihedral angles of these copolymers sug-gests a clear departure from extensive and continuous planarity in the polymer films with the most distorted being the PNDI-SVS copolymer. Physical properties, including optical bandgap, and HOMO and LUMO energetic levels, are summarized in Table 1.
2.2. Structural Characterization of Thin Films
In order to investigate the microstructure of the PNDI-SVS thin films, we performed GIWAXS on a series of samples annealed for 20 min at different temperatures up to 250 °C and for a sample annealed for 24 h at 130 °C. The 2D GIWAXS patterns are shown in Figure 1.
Increased annealing temperature decreases the orienta-tional disorder within the films, making the peaks narrower azimuthally. While up to 150 °C the films show largely liquid crystallinity with scattering peaks only clearly in-plane and out-of-plane, for T ≥ 180 °C there is clear evidence for 3D crystal-linity, with several mixed-index peaks (away from the vertical and horizontal axes) evident.
For T ≤ 210 °C and an annealing duration of 20 min, the 2D scattering patterns indicate a generally face-on orientation with
Table 1. Physical properties of PNDI-SVS.
Polymer Yield [%] Mn/Mw [kDa] λabs [nm] Td [°C] LUMOa) [eV] HOMOb) [eV] Eg [eV]
Solution Film
PNDI-SVS 97 80/168 403, 711 416, 748 387 −3.98 −5.29 1.31
a)E
full p
aper
alkyl stacking lamella peaks largely in-plane and π-stacking peaks predominantly out-of-plane. At both higher temperatures (250 °C) and longer annealing times, a clear out of plane alkyl stacking peak, dim in the other scattering patterns, becomes quite evident, indicating an edge-on realignment in the film.
To quantify the overall orientation of crystallites within the thin films, we calculated Herman’s orientational parameter (S) using previously established methods.[70] A value of S = 1
indicates a film consisting of perfectly aligned edge-on crystals, while a value of −0.5 indicates perfect face-on stacking within the film. Up to 180 °C at 20 min of annealing, S progressively decreases with increasing temperature from −0.3 to −0.4, indi-cating mostly face-on crystalline orientation and a narrowing of the distribution width, matching the narrowing polar distribu-tion we see in the 2D scattering patterns. At 210 °C, however,
S increases to −0.2 and finally switches to a positive value of
0.5 at 250 °C, indicating a significant edge-on component at 210 °C, which becomes the predominant stacking orientation within the film at 250 °C. Similar edge-on predominance is seen at longer annealing times in the film annealed for 24 h at 130 °C, indicating that the edge-on orientation is energetically
preferred even at lower temperatures, which may require more time for large-scale realignment and propagation throughout the film owing to lower mobility.
To examine the crystalline parameters more carefully, in-plane and out-of-in-plane sector profiles were calculated, as shown in Figure 2.
The three unit cell directions are identified as an alkyl stacking from backbone to backbone across the alkyl side chains (≈2.6 nm), from backbone to backbone in a π-stack (0.41 nm), and from one monomer to the next along the back-bone (1.6 nm), which are labeled in the 1D sector profiles in Figure 2. Upon annealing initially to 120 °C for 20 min, we see a general loss of crystallinity and a decrease of coherence length, indicating that at this low temperature (and short annealing time), the crystals that were formed upon casting are disrupted slightly. Upon annealing at 180 °C for 20 min, we can see that in all unit cell directions, both the coherence length and the relative crystallinity increase considerably. Upon annealing at 210 °C for 20 min, while the crystallinity remains high and even increases, the coherence length of those crystals decreases slightly, as this is the temperature at
Figure 1. 2D GIWAXS patterns plotted versus in-plane and out-of-plane momentum transfer (Qxy and Qz, respectively) for a) as cast, b) 120 °C 20 min,
c) 150 °C 20 min, d) 180 °C 20 min, e) 210 °C 20 min, f) 250 °C 20 min, and g) 130 °C 24 h. h) Herman’s orientational parameter (S) calculated from the polar distribution of the second order alkyl stacking peak, plotted versus annealing temperature.
full p
aper
which the molecular packing orientation was seen to begin to change. Note that because of the large-scale realignment of the film at 250 °C and for the longer anneal at 130 °C, many of the crystalline planes that contribute to each peak are scattering to different directions, making it quite difficult to compare the peak area and spacing accurately to those at the lower temperatures/shorter annealing times. We can however see that the coherence length remains at about the same level, and there is no indication of a drastically dif-ferent unit cell.
Subtle changes to the unit cell are evident upon annealing, with the alkyl stacking distance and backbone repeat distance increasing up to 180 °C, which indicates a straightening of the backbone and alkyl side chains. Meanwhile, the π-stacking dis-tance remains largely unchanged. Upon annealing to 210 °C, the unit cell dimensions reduce once again. We speculate that the large crystals with straightened backbone obtained at 180 °C are no longer stable at 210 °C, as the film begins to reorient. At higher temperatures, crystallinity is high once again, indicating that after reorientation, crystals can again grow freely.
Figure 2. 1D 15° sector profiles calculated from the 2D GIWAXS patterns. a) Scattering intensity versus in-plane momentum transfer Qxy. b)
Scat-tering intensity versus out-of-plane momentum transfer Qz. Peak fitting results are displayed for c) the alkyl stacking peak, d) the π-stacking peak, and
full p
aper
Long-duration annealing at low temperature (130 °C, 24 h), as previously noted, leads to similar crystalline structures and orientations as those found at higher temperatures for shorter times (250 °C, 20 min), revealing no evidence of a phase tran-sition for T ≤ 250 °C. The one difference in the crystalline structure is that the coherence lengths along all directions are much higher in the high-temperature annealed film, indicating that the quality and size of the crystallites are improved when using high-temperature short-duration thermal treatments rather than low-temperature long-duration treatments.
To investigate the overall alignment of the backbone, both at the surface and within the bulk of the polymer, near edge X-ray absorption fine structure (NEXAFS) spectroscopy was collected on films where a largely face-on bulk orientation was demon-strated by GIWAXS measurements, i.e., as-cast, 120 °C 20 min, 150 °C 20 min, 180 °C 20 min, and 210 °C 20 min. The total electron yield (TEY) results are presented in Figure 3, along with the tilt angles determined through TEY, partial electron yield (PEY), and fluorescence yield (FY).
PEY is the most surface-sensitive of the techniques, meas-uring the upper 1–3 nm of the film. TEY is also surface-sen-sitive, although it is sensitive to the upper 2–5 nm of the film. FY is bulk sensitive, measuring the entirety of the thin film. All tilt angles are calculated from a peak fit of structure at lower energies than 286 eV, which is in the π* manifold, corresponding to transition dipole moments normal to the face of a conjugated core, along the direction of the carbon π orbitals. The NEXAFS tilt angles for the as-cast films show a marked difference between the surface layer, as revealed by PEY and TEY, and the bulk of the thin films, as revealed by FY. At the surface, tilt angles are consistently more edge-on than within the bulk. Annealing at 150 and 180 °C produces the highest edge-on orientation, with the highest tilt angle being 63.6° at 150 °C. Further annealing to 210 °C decreases the surface tilt angle back toward 55°. The bulk of the films maintain a highly face-on orientation, with a general
decrease in this orientation upon annealing consistent with the GIWAXS results.
Although beyond the scope of this paper, we want to note here that through careful angle-dependent GIWAXS measurements and analysis, a crystalline self-stratification effect was observed in the film annealed to 210 °C for 20 min. This resulted in a 9 nm edge on-crystalline region lying on top of a 63 nm face-on crystalline regiface-on. This surface orientatiface-onal segregatiface-on matches the overall orientational difference we measure in NEXAFS, and provides evidence that this surface region is crys-talline. Details and further discussion are reported in a separate work.[71] AFM investigation of PNDI-SVS films revealed a
tex-tured topography composed of elongated structures, common to both as-cast and annealed films (210 °C, 20 min) (Figure 4). Such topographic features are usually observed in the parent polymer P(NDI2OD-T2) and they have been ascribed to self-assembling phenomena occurring during film deposition and solidification and just initiated by aggregates present in solu-tion;[3] this correspondence with P(NDI2OD-T2) morphology is
also reflected by GIWAXS, revealing a mostly face-on stacking in the bulk, and by NEXAFS, evidencing a pronounced edge-on molecular arrangement at the surface.[72] In the case of the
as-cast film (Figure 4a) (Rrms= 0.7 nm), topographic features
are many tens of nm wide (up to 40 nm), while their length is inaccessible, with the elongated structures being mostly inter-twined and interconnected to each other; however, segments up to 500 nm long with coherent directionality can be observed. From the section analysis of those nanostructured segments that are less woven and more clearly stacked (Figure 4b), thick-ness values of ≈2.3–2.9 nm were extracted, thus in good agree-ment with a mostly edge-on orientation. The sample annealed at 210 °C (Figure 4c) displays a slightly rougher surface (Rrms= 0.9 nm), likely reflecting the enhanced crystallinity of
the film, in which the elongated microstructures appear much more fused one into another. The section analysis of such a top layer is accessible through discontinuity regions (Figure 4d),
Figure 3. The NEXAFS results of PNDI-SVS films upon annealing. TEY spectra across the Carbon 1S absorption edge are shown for a) as cast and
full p
aper
revealing its monomolecular nature and slightly inferior thick-ness values compared to the nanostructures observed in the topography of the as-spun sample (≈2.0–2.6 nm), which is again in agreement with the reduction of the tilt angle pre-dicted by NEXAFS on the same sample.
It is worth highlighting that a clear discrepancy between bulk (observed by GIWAXS and FY NEXAFS) and top surface (observed by PEY, TEY NEXAFS, and AFM) structural evolu-tion with temperature was observed: the first one showing a clear transition from mostly face-on to largely edge-on crystal-lite orientation, the second one displaying a majority of edge-on molecular orientatiedge-on already in as-cast films and maximum tilt angle values for films annealed around 150–180 °C. There-fore, it is possible that there is a subtle surface reconstruction with annealing that is not discernible with GIWAXS and whose structure is hard to elucidate with NEXAFS, but which is clearly suggested by topographic investigation.
2.3. Electrical Characterization of FET Devices
We fabricated top-gate, bottom-contact PNDI-SVS FETs using poly(methyl methacrylate) (PMMA, ε = 3.6) (Figure 5 and Table 2) as the dielectric layer. All the FETs exhibited typical ambi-polar characteristics, with high electron mobility and low hole mobility. Generally, an incremental enhancement of the source to drain electron current (IDS) with temperature is observed
when the semiconductor undergoes 20 min thermal annealing in nitrogen atmosphere (Figure 6a): correspondently, we have
extracted increasing electron mobility values (μe) at VGS =
VDS = 60 V, going from the pristine (as cast) film (μe ≈
0.1 cm2 V−1 s−1) up to films annealed at T = 150 °C, reaching
mobility values eventually superior to 2.0 cm2 V−1 s−1. Annealing
at temperatures higher than 150 °C does not provide substantial improvement of the OFETs performances, however a reduction of threshold voltages can be observed, allowing an average μe of
2.4 cm2 V−1 s−1 to be achieved (T = 250 °C). The introduction
of a very thin injection/doping layer of Cs2CO3[73,74] between
the source and drain electrodes and PNDI-SVS film results in μe ≈ 0.5 cm2 V−1 s−1 using as cast films (Figure S6, Supporting
Information); thus, already five times higher mobilities can be achieved without necessarily taking advantage of a thermally induced structural reorganization of the as-cast polymeric films.
Another major point of concern for the practical use of poly-meric OFETs is the ambient stability, especially in the case of
n-type polymers.[75] Device instabilities are usually observed in
the form of hysteretic transfer curves, accompanied by a gradual decrease in the channel current under constant bias conditions, when operated in ambient air without encapsulation. This latter occurs as a consequence of a shift in the threshold voltage (Vth)
and also of a drastic reduction of the charge carrier mobility. To check the ambient and bias stress stability of PNDI-SVS FETs, we measured the FET characteristics of as-cast (Figure S7, Supporting Information) and 210 °C annealed films under continuous switch on and off in air at a constant gate voltage (VGS = 60 V) for more than 8 h (Figure 5c) and under continuous
biasing at constant VD = 30 V and VG = 60 V for 2 h (Figure 5d,
Supporting Information). Both the on and off currents remained
Figure 4. a,c) AFM image images and b,d) section analysis of superficial nanostructures of (a,b) PNDI-SVS as-cast film and (c,d) film annealed at
full p
aper
substantially constant upon on–off cycling (Figure S7a, Sup-porting Information and Figure 5c) and a constant biasing in air resulted in just a 3% reduction of the on current after 2 h in the case of 210 °C annealing (Figure 5d) and an 18% reduc-tion in the as cast film (Figure S7b, Supporting Informareduc-tion), suggesting in both cases a stable device threshold voltage and quite unmodified transport characteristics throughout both the stress tests. The stability of PNDI-SVS OFET is thus similar to other air-stable NDI copolymers with extended donor unit, such as PNDI-TVT, and apparently improved when compared to the parent PNDI2OD-T2 with bithiophene donor; while the rela-tively low LUMO energetic value (−3.98 eV) of the polymer may
have resulted in reduced electron transfer phenomena, which can initiate NDI degradation,[76] the comparison between the
as-cast and the annealed films highlights the contribution of the superior crystallinity in improving electrical stability in air, likely owing to a more efficient protection of the polymeric charged
Table 2. Device parameters of PNDI-SVS OFETs with various annealing
conditions (W/L = 1 mm/30 μm). Annealing temperature (TA) [°C]/[min] S/D electrode μe,sat [cm2 V−1 s−1] V[V]e,Th [V decS.S. −1] Pristine Au 0.1(± 0.02) 12.2 (± 1.1) 20.0 (± 0.7) Cs2CO3/Au 0.5 (± 0.15) 7.1 (± 1.5) 4.3 (± 0.6) 120/20 Au 0.3 (± 0.08) 19.9 (± 1.1) 16.26 (± 1.2) 150/20 Au 1.9 (± 0.22) 18.2 (± 0.7) 16.1 (± 1.1) 180/20 Au 1.9 (± 0.54) 17.1 (± 0.7) 12.3 (± 0.6) 210/20 Au 2.2 (± 1.01) 16.4 (± 0.8) 14.5 (± 1.3) 250/20 Au 2.4 (± 0.39) 14.6 (± 1.0) 10.0 (± 0.2)
Figure 6. The DOS of PNDI2OD-T2 (black) and PNDI-SVS (red) annealed
at 120 °C (black). The shaded areas represent the filling of the DOS at
VG = 60 V.
Figure 5. a) Transfer characteristics of PNDI-SVS OFETs (L/W = 30 μm/1 mm) fabricated as cast (black) and annealed at 250 °C (orange) and b) square root drain current curves. c) Programming cycling test with continuous VD ( = 60 V) bias and switched VG bias (off at 12 V, on at 60 V).
full p
aper
core from the contact with oxygen and water, operated through the branched hydrophobic side chains.
2.4. Electronic Structure
To assess the electronic structure of PNDI-SVS, we adopted a methodology to extract the density of states (DOS) using Metal Insulator Semiconductors (MIS) structures.
Maddalena et al. recently developed a tool of general appli-cability which is based on populating and probing the DOS in the quasi-static regime by means of capacitive coupling in MIS structures.[62] The main advantage with respect to other
elec-trical techniques lies in the fact that by operating the MIS in the quasi-static regime carrier transport is not involved; hence, the DOS can be investigated without making hypotheses or assumptions on the charge transport model.
Experimental capacitance–voltage (CV) characteristics of MIS devices are fitted assuming that the DOS can be modeled as a superposition of Gaussian functions, each characterized by its own width (σ) and number of states (N0).[62] For films
sub-jected to 18 h of thermal annealing at 130 °C after deposition, our best fit yields an effective DOS composed of two Gaussian functions, a main one and a secondary, smaller, intragap one. For the main one, the center was set at the polymer LUMO level, N0,1 was set to 1021 cm−3, and the best fit was obtained
using σ1= 92 meV. For the secondary Gaussian, best fits were
obtained shifting its center by 0.1 eV with respect to the main one and using N0,2= 1018 cm−3 and σ2= 161 meV. The fitted
DOS is plotted in Figure 6, where the shaded areas represent the filling of the DOS at a gate voltage of 60 V. The total concen-tration of charge carriers at 60 V is of 1.6 × 1019 cm−3. The same
method employed to extract the DOS of the parent polymer PNDI2OD-T2, results in a single Gaussian function with a width of 78 meV,[62] much narrower than PNDI-SVS films
opti-mized for transport.
The secondary Gaussian of PNDI-SVS might be related to intragap, low mobility states. As for the main PNDI-SVS Gaussian, it might be surprising that its width is larger than PNDI2OD-T2, given the fact that PNDI-SVS carrier mobility is higher (a DOS comparison is shown in Figure 6). To reconcile these aspects, we can speculate that PNDI-SVS DOS is spatially correlated.[77] In fact, as far as carrier mobility is concerned,
spa-tial correlation can counterbalance energetic disorder because it makes the energy level landscape smoother with respect to the uncorrelated case. Hence, we show that a broadening of the effective DOS does not necessarily imply a poorer charge transport, as may be assumed, and a strong influence of mor-phological factors should necessarily be taken into account to properly model the transport properties of a polymeric channel.
2.5. Polarized Charge Modulation Microscopy
From the electrical characterization of the FETs reported in the previous section, a major point regarding the functional microstructure–properties nexus arises: low-temperature annealed films (150 °C) display electron transport proper-ties comparable to films that underwent strong structural
rearrangement through higher-temperature thermal annealing. Thus, a correlation between structure and mobility (as calculated from a FET) or DOS (as extracted from MIS structures) cannot be easily traced. When trying to establish such a nexus, we have first to take into account that, in FETs in full accumulation, charges probe only a few nm-thick semi-conductor layer at the interface with the dielectric, and this critical region may display a different microstructure from the bulk.[72] Moreover, charge transport in organic semiconductors
not only depends on features like crystallinity, but, especially in films with a consistent amorphous-like phase, important factors are also film interconnectivity[1,78] and degree of
non-crystalline orientational order,[79] as recently evidenced for
example for PNDI2OD-T2.[3,80] Therefore, typical bulk average
structural characterization of the ordered phase, as obtained by GIWAXS, has to be completed with more specific, channel sensitive investigations. For the PNDI-SVS copolymer investi-gated in this work, AFM surface analysis and surface-sensitive NEXAFS reveal that the packing motif at the surface of the film is prevalently edge-on, compared to the prevalently face-on packing of the bulk. Moreover, angle-dependent GIWAXS measurements have highlighted a stratification effect resulting in a 9 nm edge-on crystalline region lying on top of a 63 nm face-on crystalline region.[71] It therefore becomes essential to
selectively analyze the functional morphology of the accumu-lation channel of films in actual devices and to directly image the charge distribution along the channels of working devices. We previously showed that charge modulation microscopy (CMM)[61] probes only those mobile charges accumulated at
the gate dielectric. Therefore, only charged conjugated seg-ments which can be populated and depopulated at the gate bias frequency modulation can be imaged. Furthermore, when a polarized light is used for the probing beam (p-CMM), the angular mapping of the in-plane component of the transi-tion dipole moment (TDM) of the charge-induced optical tran-sitions is also possible. In previous work it has been shown that the TDM can be also related to the preferential in-plane alignment of the polymer backbone.[60]
Before proceeding with the charge modulation mapping, we acquired the charge modulation spectra (CMS) of PNDI-SVS (Figure S8, Supporting Information), which is overall red-shifted with respect to the one of PNDI2OD-T2, reflecting the smaller band-gap. Both spectra show two peaks in the ground state absorption bleaching region (ΔT/T > 0), and one polaronic absorption peak (ΔT/T < 0). Such features have been previously assigned.[39]
According to the observed CMS spectra, we performed the charge modulation mapping by probing the samples at 745 nm in correspondence with the bleaching signal of the main absorption peak. The CMM data acquired on PNDI-SVS FETs annealed at 120° (20 min), 250° (20 min) and as-cast in the presence of Cs2CO3 as the injection layer, are presented in Figure 7. The orientation maps found per each device treat-ment (Figure 7a–c) represent the average TDM angular orienta-tion of the bleached ground state absorporienta-tion, i.e., the average orientation along the channel area of polymer backbones where mobile charge resides. We also report the degree of order (DO) maps in Figure 7d–f. The DO corresponds to the ratio between the light polarization-dependent signal and the total charge
full p
aper
modulation signal, indicating the fraction of the signal arising from aligned TDM.[60]
In all maps, clear orientational domains of the optical dipoles can be observed, likely sustained by the fibrillar-like supra-molecular morphology observed with AFM at the top surface of the films. According to the maps, charge is probing orientational domains with a limited spatial extent (hundreds of nanome-ters), with a largely variable orientation in the mapped area and with only a slight orientational coherence in the long-range. Importantly, from each map, average degree of order values of 48%, 44%, and 51% were calculated for 120 °C, as-cast on Cs2CO3, and 250 °C samples, respectively. As a comparison, in
the case of the parent PNDI2OD-T2, a clear correlation between transport properties and orientational order was observed with p-CMM:[60] when a high degree (≈100%) of order is achieved,
mobility superior to 1 cm2 V−1 s−1 is measured, while cases of
lower degree of order, similar to that observed in PNDI-SVS, lead to mobilities of ≈0.1 cm2 V−1 s−1. In the case of PNDI-SVS,
maps are found to be quite invariant with thermal treatments both in terms of orientation and degree of order. This is clear evidence that thermal treatment only is not active on the ori-entational order of the functional phase of the polymer and that the strong bulk morphological rearrangements observed with GIWAXS is not reflected in the FET channel orientational morphology of PNDI-SVS films. Conversely, CMM maps are in agreement with the quite invariant orientational arrangement of the molecules in the top layers of the film observed with NEXAFS measurements.
Despite the similar CMM maps, the investigated films have shown quite different charge transport properties, the film annealed at high temperature displaying higher mobility values than the as-cast and the 120 °C annealed films (see Table 2).
Figure 7. CMM maps of the angle orientation of the charge transition dipole moment measured on PNDI-SVS-FET (Vg = 40 V; Vpp = 40 V; modulation
frequency 989 Hz): a) FET with semiconductor film annealed at 250 °C, 20 min; b) annealed at 120 °C, 20 min; c) as-cast with Cs salt. On the right the degree of order maps corresponding respectively to d) 250 °C, e) 120 °C, f) as-cast with Cs salt. The source and drain electrode are at the left and right sight of the scan area. The maps are reproducible over different sampling areas.
full p
aper
Part of the increase can be attributed to a more efficient charge injection, since a five times higher mobility is observed already in as-cast samples with the insertion of a thin Cs2CO3 charge
injection layer. However, there is a further increase of approxi-mately a factor of four for temperatures >150 °C, which has to derive from improved transport properties. Since the micro-metric orientational domains displayed by CMM are already present in low-temperature films, the bottleneck for charge transport cannot be found on such length scale. The improved transport therefore has to derive from a finer microstructural evolution, where already for annealing temperatures superior to 120 °C, a packing reorganization favors a more efficient interchain transport and domains interconnectivity. Such evolu-tion is evidenced by the change in molecular tilt angle observed with PEY NEXAFS starting from 150 °C (Figure 3) and by the film topography (Figure 4), evolving with thermal annealing from a clear fibrillar structure toward a much more intercon-nected layer. It may also be speculated that a similar improved coherence length in each crystalline direction with tempera-ture, as measured on average in the film with GIWAXS, may also undergo at the top surface of the film.
It is worth noting that in the case of PNDI-SVS, high mobility values in the range of 2 cm2 V−1 s−1 are achieved with a lower
degree of orientational order with respect to PNDI2OD-T2, thus suggesting that substitution of thiophenes with electron-rich selenophenes results in more effective intermolecular charge transfer and that there may be room for further improvement of the electrical properties of PNDI-SVS through the enhance-ment of the film long-range orientational order.
3. Conclusions
We have reported the synthesis, thin film structure, and elec-tronic properties of a novel electron-transporting NDI-copolymer, PNDI-SVS featuring a selenophene-vinyl-selenophene donor moiety, showing a reduced energy bandgap of 1.31 eV and a slightly lower lying LUMO level (−3.98 eV) with respect to the extensively studied PNDI2OD-T2 reference copolymer. PNDI-SVS, enables high-performance solution-processed OFETs, with optimized devices showing electron mobility of 2.4 cm2 V−1 s−1 and a promising stability for ambient air
opera-tion, prior to any specific encapsulaopera-tion, which is in any case required for applications. Its DOS is interestingly character-ized by the superposition of two Gaussian functions, one shifted 0.1 eV with respect to the other yielding a relatively broad distribution, where the central and dominant one has a total number of states N0,1 = 1021 cm−3 and width σ1 = 92 meV,
and the smaller one N0,2= 1018 cm−3 and σ2= 161 meV. While
also the central Gaussian is broader than what can be extracted for the reference copolymer PNDI2OD-T2, electron mobility is conversely much improved, suggesting a possible spatial cor-relation in the DOS.
Through GIWAXS measurements, we evidenced a marked morphology evolution with annealing temperature treatments of PNDI-SVS thin films toward a more crystalline phase, where for temperatures above 180 °C a remarkable 3D crystal struc-ture arises with the presence of several mixed index peaks. Molecular packing within the bulk is prevalently face-on up to
210 °C, while a clear realignment of the film toward edge-on packing occurs for higher temperatures. No apparent phase transition is present for temperatures below 250 °C, as we have evidenced that similar packing and morphology can be accessed both at high temperatures for a short processing time (250 °C, 20 min), and at low temperatures for a long time (130 °C, 24 h), where the main difference is observed only in the coherence lengths along all directions, indicating the strongly improved quality and size of crystallites for higher temperatures.
The FETs electrical characteristics show mobility of ≈2 cm2 V−1 s−1 at an annealing temperature of 150 °C, clearly
deviating from a simplistic explanation where the trans-port properties can be directly correlated with bulk sensitive GIWAXS measurements.
NEXAFS analysis of the topmost molecular layers, where charge accumulation mainly takes place, reveals a tendency toward edge-on molecular packing, even at low temperatures and in the as-cast film, consistently with AFM topography, indi-cating a fibrillar-like supramolecular morphology. This struc-tural evidence is complemented by charge-modulation micro-scopy measurements, which reveal that mobile charge probe hundreds of nanometers wide orientational domains. Such domains are found to be very similar for devices that under-went very different processes, and in particular for those based on as-cast films on Cs2CO3, for films annealed at 120 °C and at
250 °C, all displaying a limited orientational order around 50%. On the one hand, such evidence suggests a possible further improvement of mobility with improved directional order, on the other, it indicates that improved transport properties above 150 °C are linked to a microscopic structural evolution at the top surface, as evidenced in NEXAFS and AFM, improving film interconnectivity thanks to a more efficient π-stacking within a fixed extension and order of orientational domains.
The achievement of high mobility with relatively low pro-cessing temperatures is a critical aspect to enable applications in flexible, large-area plastic electronics. We rationalize the good transport properties achieved at low temperature with the pres-ence of a segregated polymer phase forming at the liquid–air interface, following different formation dynamics with respect to the bulk, which mostly presents an edge-on packing motif and where charge follows orientational domains rather insensi-tive to annealing conditions.
Further development in the field of high-performing semi-conducting polymer FETs will be strongly influenced by the correct evaluation of the solid-state morphology and electronic properties of the sole fraction of molecules, the so-called “func-tional morphology”, contributing to transport in FET devices, especially in polymer systems with highly localized carriers and elongated, fibril supramolecular structures, favoring the forma-tion of direcforma-tional domains.
4. Experimental Section
Materials and Characterization: All starting materials were purchased
from Sigma-Aldrich
(South Korea), TCI, and Alfa Aesar, and were used without further purification. Pd catalyst was purchased from Amicore. 1H NMR
spectra were recorded using a Bruker Avance 300 MHz and Bruker 500 FT-NMR spectrometer; chemical shifts (ppm) were reported with
full p
aper
tetramethylsilane as an internal standard. Fourier transform (FT)-IR spectra were recorded using a Bruker IFS66 spectrometer. TGA was performed under N2 using a TA instrument 2050 TG analyzer. DSC
was conducted under N2 using a TA instrument DSC Q10, and both
samples were heated at a rate of 10 °C min−1. UV–vis spectra were
measured using a Shimadzu UV-1065PC UV–vis spectrophotometer. The electrochemical properties of the materials were measured by cyclic voltammetry (CV) using Epsilon C3 in a 0.1 m solution of
tetrabutyl ammonium perchlorate (TBAP) in chloroform. Molecular weights and polydispersities of the copolymers were determined by ACQUITY APC System with polystyrene standard calibration (solvent: chloroform).
Synthesis of 4,9-Dibromo-2,7-bis(2-Decyltetradecyl)benzo[lmn] [3,8]-Phenanthroline-1,3,6,8(2H,7H)-Tetraone (NDI-Br2) (1): The reaction
followed the reported method.[69] 4,9-Dibromoisochromeno[6,5,4-def]
isochromene-1,3,6,8-tetraone (NDA-Br2) (10.00 g, 23.47 mmol) and
2-decyltetradecan-1-amine (20.75 g, 58.68 mmol) were added to 200 mL of acetic acid. The reaction mixture was refluxed under N2 for 1 h. Then
the reaction mixture was poured onto water and the resulting precipitate was filtered and washed with methanol. The crude product was purified via column chromatography over silica gel (eluent: dichloromethane/ hexane = 1/1). NDI-Br2 was obtained as a light yellow solid. Yield:
40% (10.36 g). 1H NMR (300 MHz, CDCl 3, δ): 9.01 (s, 2H), 4.16–4.14 (t, 4H), 2.15–2.00 (m, 2H), 1.27 (m, 80H), 0.92–0.88 (t, 12H); 13C NMR (125 MHz, CDCl3, δ): 163.81, 163.65, 141.79, 131.00, 130.38, 127.92, 126.71, 48.08, 39.10, 34.56, 34.55, 34.18, 32.66, 32.33, 32.31, 32.30, 32.28, 32.26, 32.23, 32.00, 31.98, 28.97, 25.34, 25.33, 16.76; IR (KBr, cm−1): ν = 3048, 2926, 2846, 1710, 1655.
Synthesis of Selenophene-2-Carbaldehyde (2): The reaction followed
the reported method.[46] The crude product was purified via column
chromatography over silica gel (eluent: dichloromethane/hexane = 1/9). Compound 2 was obtained as a yellow oil. Yield: 70% (8.49 g). 1H NMR
(300 MHz, CDCl3, δ): 9.80 (s, 1H), 8.49 (d, 1H), 8.01 (dd, 1H), 7.47
(dd, 1H); 13C NMR (75 MHz, CDCl
3, δ): 184.36, 150.37, 141.12, 139.56,
130.86.
Synthesis of (E)-1,2-Di(selenophen-2-yl)Ethene (3): The reaction
followed the reported method.[63] The crude product was purified via
column chromatography over silica gel (eluent: dichloromethane/hexane = 1/5). Compound 3 was obtained as a yellow powder. Yield: 53% (4.06 g). 1H NMR (300 MHz, CDCl 3, δ): 7.90–7.88 (dd, 2H), 7.25–7.23 (m, 4H), 7.05 (s, 2H). 13C NMR (75 MHz, CD 2Cl2, δ): 148.29, 130.19, 129.03, 128.95, 125.04. Synthesis of (E)-1,2-Bis(5-(trimethylstannyl)Selenophen-2-yl)Ethene (4): The reaction followed the reported method.[63] The crude solid was
recrystallized from methanol to obtain (E)-1,2-bis(5-(trimethylstannyl)-selenophen-2-yl)ethene as yellow needles. Yield: 74% (2.02 g). 1H NMR
(300 MHz, CD2Cl2, δ): 7.96–7.95 (m, 2H), 7.46–7.44 (d, 2H), 7.04–7.00 (dd, 2H), 0.37 (s, 18H). 13C NMR (125 MHz, CD 2Cl2, δ): 153.47, 145.03, 138.50, 129.91, 125.22, −8.34. Synthesis of Poly[(E)-2,7-Bis(2-decyltetradecyl)-4-methyl-9-(5-(2-(5-methylselenophen-2-yl)vinyl)Selenophen-2-yl)Benzo[lmn][3,8]- Phenanthroline-1,3,6,8(2H,7H)-Tetraone] (PNDI-SVS): The polymer
was prepared from a palladium-catalyzed Stille coupling reaction. 4,9-Dibromo-2,7-bis(2-decyltetradecyl)benzo[lmn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone (NDI-Br2) (0.50 g, 0.46 mmol) and
(E)-1,2-bis(5-(trimethylstannyl)selenophen-2-yl)ethene (0.28 g, 0.46 mmol) were dissolved in dry chlorobenzene (7.5 mL). After degassing with nitrogen for 1 h, Pd2(dba)3 (8 mg) and P(o-Tol)3 (11 mg) were added
to the mixture and stirred for 48 h at 110 °C. 2-Bromothiophene and tributyl(thiophen-2-yl)stannane were added successively at a time interval of 6 h to end-cap the end groups. The polymer was precipitated in methanol. The crude polymer was collected by filtration and then purified by Soxhlet extractions with methanol, acetone, hexane, toluene, and chloroform, successively. The final product was obtained by precipitation in methanol and drying in vacuo. PNDI-SVS was obtained as a dark green solid. Yield: 97% (0.54 g). 1H NMR (500 MHz, CDCl
3,
δ): 8.80–8.46 (br, 2H), 7.93–7.20 (br, 6H), 4.12–4.10 (br, 4H), 1.98–1.87 (m, 2H), 1.25 (br, 80H), 0.87–0.86 (m, 12H).
Field-Effect Transistor Fabrication and Characterization: The Au/Ni
(15 nm/3 nm thick) patterns used for the source and drain electrodes were fabricated using a conventional photolithography procedure on Corning Eagle 2000 glass substrates. Substrates were cleaned sequentially in an ultrasonic bath with de-ionized water, acetone, and iso-propanol and then baked at 110 °C. Au source and drain electrodes of some devices were treated with Cs2CO3 as an electron injection layer.
The Cs2CO3 (Sigma-Aldrich) was dissolved in 2-ethoxyethanol and
spin-coated in a N2-purged glove box onto a patterned glass substrate with
Au bottom contact electrodes. The substrates were thermally annealed at 120 °C for 30 min in an N2-filled glove box. The conjugated polymer
PNDI-SVS was dissolved in anhydrous chlorobenzene solvent to obtain a 12 mg mL−1 solution. The semiconductor solutions were filtered
with a 0.2 μm polytetrafluoroethylene syringe filter and spin coated at 2000 rpm for 60s in an N2-filled glove box. The spin-coated PNDI-SVS
films were thermally annealed at various temperatures from 120 to 250 °C for 20 min in the N2-purged glove box. PMMA (MW= 120 kD)
was used as the dielectric without further purifications. PMMA was dissolved in n-butylacetate, spin-coated on the semiconducting polymer film and baked at 80 °C to remove residual solvent under N2-purged
atmosphere. The transistors were completed by depositing the top-gate electrodes (Al) via thermal evaporation using a metal shadow mask. Devices used in the ambient stability tests were placed in a controlled humidity atmosphere on the completion of the fabrication process. All FETs electrical characteristics were measured using HP 4156A in an N2
-filled glove box. The μFET and VTh were calculated at the saturation region
using gradual channel approximation equation.[81]
Preparation of Metal–Insulator–Semiconductor Capacitors for Determination of the Density of States: The MIS-capacitors were prepared
with gold as the bottom injecting contact, defined by photolithography. PNDI-SVS was spin-coated from anhydrous chlorobenzene and annealed at 120 °C for 20 min. The PMMA dielectric was spin-coated on top of the semiconductor and aluminum was thermally deposited to form the gate electrode. Finally, the devices were annealed overnight at 130 °C. The CV characteristics of the MIS-capacitors were measured using an Agilent E4980A Precision LCR Meter. CV curves were fitted using a custom software developed to solve the nonlinear Poisson equation and to calculate the capacitance of the device at each applied voltage.[62]
GIWAXS measurements: GIWAXS measurements were carried out
with a Dectris Pilatus 1 m detector. 11 or 18 keV (for the 250 and 130 °C
24 h annealed samples) highly collimated photons were aligned normal to the sample by using a photodiode. Angular steps of 0.01° were taken near the critical angle, which was determined as the angle of maximum scattered intensity. Each 2D scatter pattern was a result of a total of 3 s of exposure. Three 1 s exposures were taken at different detector positions to fill in the gaps between modules in the detector, and combined in the software. Correction of data onto momentum transfer axes and sector profiles was done using an altered version of NIKA.[82]
Peak fitting was done using least squares multipeak fitting within Igor Pro. By identifying the scattering peaks, and fitting them with a log cubic empirical background function, we can monitor the crystallinity in the three major unit cell directions. Peak area is proportional to the number of molecular planes that participate in crystalline stacking and so the relative crystallinity within the thin film, while the peak width is an indication of the coherence length of that crystallinity and represents the size and quality of those crystals, and finally the location of the peak gives the d-spacing or the distance from one unit cell to the next along that direction.
NEXAFS measurements: NEXAFS measurements were conducted
using X-rays produced by an elliptically polarized undulator and a 1200 lines/mm grating, producing highly polarized X-rays with a bandwidth of ≈0.1 eV across the measurement range. TEY was measured by monitoring the drain current to the sample, PEY was measured with a channeltron detector with a retarding voltage of ≈200 V, and FY was measured by a multichannel plate detector with a retarding field of 2600 V. Dark levels were measured and subtracted from the data, and double normalization was done using a gold mesh corrected by a
full p
aper
photodiode.[83] Peak fitting and tilt angle calculations were done using
the Quick AS NEXAFS Tool.[84]
All tilt angles are calculated from a peak fit of structure at lower energies than 286 eV, which is in the π* manifold, corresponding to transition dipole moments normal to the face of a conjugated core, along the direction of the carbon π orbitals. Thus, a tilt angle of 90° indicates that every TDM is perfectly in plane, and so all the conjugated faces are oriented perfectly edge-on to the substrate, while a tilt angle of 0° indicates perfectly face-on orientation. Because of the uniaxial alignment of spin-coated films, an unaligned film will have an average tilt angle of ≈54.7°. It is also impossible to determine through only a tilt angle measurement the distribution of local tilt angles within the film, but only the ensemble tilt angle. Thus, any fixed dihedral angle between backbone components will result in an average tilt angle between the tilt angles of each component. As opposed to GIWAXS, NEXAFS is sensitive to all molecules, regardless of them being in a crystalline or amorphous region.
GIWAXS and NEXAFS experiments were performed on the SAXS/ WAXS[85] and Soft X-ray[86] beamlines at the Australian Synchrotron,
Victoria, Australia.
Polarized charge Modulation Microscopy: The p-CMM data were
collected with a homemade confocal microscope operating in transmission mode. The light source consisted of a supercontinuum laser (NKT Photonics, SuperK Extreme) monochromated by an acousto-optic modulator (NKT Photonics, SuperK Select) in the 500–1000 nm region with line widths between 2 and 5 nm. Laser polarization was controlled with a half-wave plate and a linear polarizer. The light was then focused on the sample with a 0.7 N.A. objective (S Plan Fluor60, Nikon) and collected by a second 0.75 N.A. objective (CFI Plan Apochromat VC 20, Nikon). The collected light was focused to the entrance of a multimodal glass fiber with a 50 μm core, acting as the confocal aperture. Detection was operated through a silicon photodetector (FDS100, Thorlabs). The signal was amplified by a transimpedance amplifier (DHPCA-100, Femto) and supplied both to a DAQ (to record the transmission signal, T) and to a lock-in amplifier (SR830 DSP, Stanford Research Systems) to retrieve the differential transmission data, ΔT. The system was run with homemade Labview software The OD values were obtained from T, as described in the Supporting Information, while the CMS data were calculated as ΔT/T. The sample was kept in an inert atmosphere by fluxing nitrogen in a homemade chamber. The gate voltage of the transistor was sinusoidally modulated at 989 Hz between 20 and 60 V with a waveform generator (3390, Keithley) amplified by a high-voltage amplifier (WMA-300, Falco Systems), while the source and drain contacts were kept at short-circuit. Signal maps were collected by raster scanning the sample at 250 ms per pixel, with a lock-in integration time of 100 ms. All data were processed with Matlab software.
Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements
M.J.S., A.L., W.-T.P. contributed equally to this work. This research was undertaken in part on the SAXS/WAXS[85] and Soft X-ray[86] beamlines at
the Australian Synchrotron, Victoria, Australia. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean Government (MSIP) (NRF-2014R1A2A2A01007159 and 2015R1A2A1A10055620), the Center for Advanced Soft-Electronics (2013M3A6A5073183 and 2013M3A6A5073172). Work performed at IIT was financially supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme “HEROIC”, Grant Agreement No. 638059. Work performed
at Monash University was supported by the Australian Research Council (DP130102616).
Received: March 4, 2016 Published online: May 18, 2016
[1] R. Noriega, J. Rivnay, K. Vandewal, F. P. V. Koch, N. Stingelin, P. Smith, M. F. Toney, A. Salleo, Nat. Mater. 2013, 12, 1038. [2] H. Sirringhaus, Adv. Mater. 2014, 26, 1319.
[3] S. G. Bucella, A. Luzio, E. Gann, L. Thomsen, C. R. McNeill, G. Pace, A. Perinot, Z. Chen, A. Facchetti, M. Caironi, Nat.
Commun. 2015, 6, 8394.
[4] C. Liu, Y. Xu, Y.-Y. Noh, Mater. Today 2015, 18, 79.
[5] D. Fazzi, M. Caironi, Phys. Chem. Chem. Phys. 2015, 17, 8573. [6] D. Natali, M. Caironi, Adv. Mater. 2012, 24, 1357.
[7] D. Khim, Y. Xu, K.-J. Baeg, M. Kang, W.-T. Park, S.-H. Lee, I.-B. Kim, J. Kim, D.-Y. Kim, C. Liu, Y.-Y. Noh, Adv. Mater. 2016, 28, 518. [8] S. R. Forrest, Nature 2004, 428, 911.
[9] R. R. Søndergaard, M. Hösel, F. C. Krebs, J. Polym. Sci., Part B:
Polym. Phys. 2013, 51, 16.
[10] S. G. Bucella, G. Nava, K. C. Vishunubhatla, M. Caironi, Org.
Elec-tron. 2013, 14, 2249.
[11] F. C. Krebs, J. Fyenbo, M. Jorgensen, J. Mater. Chem. 2010, 20, 8994. [12] G. Dell’Erba, A. Luzio, D. Natali, J. Kim, D. Khim, D.-Y. Kim,
Y.-Y. Noh, M. Caironi, Appl. Phys. Lett. 2014, 104, 153303.
[13] G. Azzellino, A. Grimoldi, M. Binda, M. Caironi, D. Natali, M. Sampietro, Adv. Mater. 2013, 25, 6829.
[14] G. Pace, A. Grimoldi, D. Natali, M. Sampietro, J. E. Coughlin, G. C. Bazan, M. Caironi, Adv. Mater. 2014, 26, 6773.
[15] K.-J. Baeg, M. Caironi, Y.-Y. Noh, Adv. Mater. 2013, 25, 4210. [16] G. Dell’Erba, A. Perinot, A. Grimoldi, D. Natali, M. Caironi,
Semi-cond. Sci. Technol. 2015, 30, 104005.
[17] C. M. Lochner, Y. Khan, A. Pierre, A. C. Arias, Nat. Commun. 2014,
5, 5745.
[18] G. H. Gelinck, A. Kumar, D. Moet, J.-L. van der Steen, U. Shafique, P. E. Malinowski, K. Myny, B. P. Rand, M. Simon, W. Rütten, A. Douglas, J. Jorritsma, P. Heremans, R. Andriessen, Org. Electron.
2013, 14, 2602.
[19] B. K. C. Kjellander, W. T. T. Smaal, K. Myny, J. Genoe, W. Dehaene, P. Heremans, G. H. Gelinck, Org. Electron. 2013, 14, 768.
[20] M. Kaltenbrunner, T. Sekitani, J. Reeder, T. Yokota, K. Kuribara, T. Tokuhara, M. Drack, R. Schwodiauer, I. Graz, S. Bauer-Gogonea, S. Bauer, T. Someya, Nature 2013, 499, 458.
[21] Y. Yamashita, F. Hinkel, T. Marszalek, W. Zajaczkowski, W. Pisula, M. Baumgarten, H. Matsui, K. Müllen, J. Takeya, Chem. Mater.
2016, 28, 420.
[22] I. Kang, H.-J. Yun, D. S. Chung, S.-K. Kwon, Y.-H. Kim, J. Am. Chem.
Soc. 2013, 135, 14896.
[23] G. Kim, S.-J. Kang, G. K. Dutta, Y.-K. Han, T. J. Shin, Y.-Y. Noh, C. Yang, J. Am. Chem. Soc. 2014, 136, 9477.
[24] J. Li, Y. Zhao, H. S. Tan, Y. Guo, C.-A. Di, G. Yu, Y. Liu, M. Lin, S. H. Lim, Y. Zhou, H. Su, B. S. Ong, Sci. Rep. 2012, 2, 754. [25] C. Luo, A. K. Kyaw, L. A. Perez, S. Patel, M. Wang, B. Grimm,
G. C. Bazan, E. J. Kramer, A. J. Heeger, Nano Lett. 2014, 14, 2764. [26] J. Yao, C. Yu, Z. Liu, H. Luo, Y. Yang, G. Zhang, D. Zhang, J. Am.
Chem. Soc. 2016, 138, 173.
[27] C. B. Nielsen, M. Turbiez, I. McCulloch, Adv. Mater. 2013, 25, 1859. [28] R. Stalder, J. Mei, K. R. Graham, L. A. Estrada, J. R. Reynolds, Chem.
Mater. 2014, 26, 664.
[29] T. Lei, J.-Y. Wang, J. Pei, Acc. Chem. Res. 2014, 47, 1117.
[30] W. Zhang, J. Smith, S. E. Watkins, R. Gysel, M. McGehee, A. Salleo, J. Kirkpatrick, S. Ashraf, T. Anthopoulos, M. Heeney, I. McCulloch,
full p
aper
[31] A. Luzio, D. Fazzi, F. Nübling, R. Matsidik, A. Straub, H. Komber, E. Giussani, S. E. Watkins, M. Barbatti, W. Thiel, E. Gann, L. Thomsen, C. R. McNeill, M. Caironi, M. Sommer, Chem. Mater.
2014, 26, 6233.
[32] H. J. Yun, S. J. Kang, Y. Xu, S. O. Kim, Y. H. Kim, Y. Y. Noh, S. K. Kwon, Adv. Mater. 2014, 26, 7300.
[33] B. Sun, W. Hong, Z. Yan, H. Aziz, Y. Li, Adv. Mater. 2014, 26, 2636. [34] A. Lv, S. R. Puniredd, J. Zhang, Z. Li, H. Zhu, W. Jiang, H. Dong,
Y. He, L. Jiang, Y. Li, W. Pisula, Q. Meng, W. Hu, Z. Wang, Adv.
Mater. 2012, 24, 2626.
[35] C. Kanimozhi, N. Yaacobi-Gross, K. W. Chou, A. Amassian, T. D. Anthopoulos, S. Patil, J. Am. Chem. Soc. 2012, 134, 16532. [36] L. Zhang, N. S. Colella, F. Liu, S. Trahan, J. K. Baral, H. H. Winter,
S. C. B. Mannsfeld, A. L. Briseno, J. Am. Chem. Soc. 2013, 135, 844. [37] H. Usta, A. Facchetti, T. J. Marks, Acc. Chem. Res. 2011, 44, 501. [38] H. Yan, Z. Chen, Y. Zheng, C. Newman, J. R. Quinn, F. Dotz,
M. Kastler, A. Facchetti, Nature 2009, 457, 679.
[39] V. D’Innocenzo, A. Luzio, A. Petrozza, D. Fazzi, M. Caironi, Adv.
Funct. Mater. 2014, 24, 5584.
[40] V. Lemaur, L. Muccioli, C. Zannoni, D. Beljonne, R. Lazzaroni, J. Cornil, Y. Olivier, Macromolecules 2013, 46, 8171.
[41] C. Caddeo, D. Fazzi, M. Caironi, A. Mattoni, J. Phys. Chem. B 2014,
118, 12556.
[42] R. Kim, P. S. K. Amegadze, I. Kang, H.-J. Yun, Y.-Y. Noh, S.-K. Kwon, Y.-H. Kim, Adv. Funct. Mater. 2013, 23, 5719.
[43] M. Heeney, W. Zhang, D. J. Crouch, M. L. Chabinyc, S. Gordeyev, R. Hamilton, S. J. Higgins, I. McCulloch, P. J. Skabara, D. Sparrowe, S. Tierney, Chem. Commun. 2007, 5061.
[44] D. J. Crouch, P. J. Skabara, J. E. Lohr, J. J. W. McDouall, M. Heeney, I. McCulloch, D. Sparrowe, M. Shkunov, S. J. Coles, P. N. Horton, M. B. Hursthouse, Chem. Mater. 2005, 17, 6567.
[45] D. Khim, W.-H. Lee, K.-J. Baeg, D.-Y. Kim, I.-N. Kang, Y.-Y. Noh,
J. Mater. Chem. 2012, 22, 12774.
[46] I. Kang, T. K. An, J. A. Hong, H. J. Yun, R. Kim, D. S. Chung, C. E. Park, Y. H. Kim, S. K. Kwon, Adv. Mater. 2013, 25, 524.
[47] M. Shahid, T. McCarthy-Ward, J. Labram, S. Rossbauer, E. B. Domingo, S. E. Watkins, N. Stingelin, T. D. Anthopoulos, M. Heeney, Chem. Sci. 2012, 3, 181.
[48] J. Lee, A. R. Han, J. Kim, Y. Kim, J. H. Oh, C. Yang, J. Am. Chem. Soc.
2012, 134, 20713.
[49] H. Kong, Y. K. Jung, N. S. Cho, I.-N. Kang, J.-H. Park, S. Cho, H.-K. Shim, Chem. Mater. 2009, 21, 2650.
[50] B. Kim, H. R. Yeom, M. H. Yun, J. Y. Kim, C. Yang, Macromolecules
2012, 45, 8658.
[51] T. Earmme, Y.-J. Hwang, N. M. Murari, S. Subramaniyan, S. A. Jenekhe, J. Am. Chem. Soc. 2013, 135, 14960.
[52] Y.-J. Hwang, N. M. Murari, S. A. Jenekhe, Polym. Chem. 2013, 4, 3187.
[53] M. Brinkmann, L. Hartmann, L. Biniek, K. Tremel, N. Kayunkid,
Macromol. Rapid Commun. 2014, 35, 9.
[54] H.-R. Tseng, H. Phan, C. Luo, M. Wang, L. A. Perez, S. N. Patel, L. Ying, E. J. Kramer, T.-Q. Nguyen, G. C. Bazan, A. J. Heeger, Adv.
Mater. 2014, 26, 2993.
[55] R. Joseph Kline, M. D. McGehee, M. F. Toney, Nat. Mater. 2006, 5, 222.
[56] K. Tremel, F. S. U. Fischer, N. Kayunkid, R. Di Pietro, R. Tkachov, A. Kiriy, D. Neher, S. Ludwigs, M. Brinkmann, Adv. Energy Mater.
2014, 4, 1301659.
[57] L. Biniek, N. Leclerc, T. Heiser, R. Bechara, M. Brinkmann,
Macro-molecules 2013, 46, 4014.
[58] R. Matsidik, H. Komber, A. Luzio, M. Caironi, M. Sommer, J. Am.
Chem. Soc. 2015, 137, 6705.
[59] N.-K. Kim, S.-Y. Jang, G. Pace, M. Caironi, W.-T. Park, D. Khim, J. Kim, D.-Y. Kim, Y.-Y. Noh, Chem. Mater. 2015, 27, 8345.
[60] N. Martino, D. Fazzi, C. Sciascia, A. Luzio, M. R. Antognazza, M. Caironi, ACS Nano 2014, 8, 5968.
[61] C. Sciascia, N. Martino, T. Schuettfort, B. Watts, G. Grancini, M. R. Antognazza, M. Zavelani-Rossi, C. R. McNeill, M. Caironi,
Adv. Mater. 2011, 23, 5086.
[62] F. Maddalena, C. de Falco, M. Caironi, D. Natali, Org. Electron.
2015, 17, 304.
[63] B. A. Jones, A. Facchetti, M. R. Wasielewski, T. J. Marks, J. Am.
Chem. Soc. 2007, 129, 15259.
[64] M. Caironi, C. Newman, J. R. Moore, D. Natali, H. Yan, A. Facchetti, H. Sirringhaus, Appl. Phys. Lett. 2010, 96, 183303.
[65] Z. Chen, Y. Zheng, H. Yan, A. Facchetti, J. Am. Chem. Soc. 2008,
131, 8.
[66] D. Fazzi, M. Caironi, C. Castiglioni, J. Am. Chem. Soc. 2011, 133, 47. [67] A. Luzio, D. Fazzi, D. Natali, E. Giussani, K.-J. Baeg, Z. Chen,
Y.-Y. Noh, A. Facchetti, M. Caironi, Adv. Funct. Mater. 2014, 24, 1151. [68] E. Giussani, D. Fazzi, L. Brambilla, M. Caironi, C. Castiglioni,
Mac-romolecules 2013, 46, 2658.
[69] R. Kim, P. S. K. Amegadze, I. Kang, H.-J. Yun, Y.-Y. Noh, S.-K. Kwon, Y.-H. Kim, Adv. Funct. Mater. 2013, 23, 5719.
[70] L. A. Perez, P. Zalar, L. Ying, K. Schmidt, M. F. Toney, T.-Q. Nguyen, G. C. Bazan, E. J. Kramer, Macromolecules 2014, 47, 1403.
[71] E. Gann, M. Caironi, Y.-Y. Noh, Y.-H. Kim, C. R. McNeill, unpublished.
[72] T. Schuettfort, L. Thomsen, C. R. McNeill, J. Am. Chem. Soc. 2013,
135, 1092.
[73] K.-J. Baeg, J. Kim, D. Khim, M. Caironi, D.-Y. Kim, I.-K. You, J. R. Quinn, A. Facchetti, Y.-Y. Noh, ACS Appl. Mater. Interfaces 2011,
3, 3205.
[74] C. Liu, J. Jang, Y. Xu, H.-J. Kim, D. Khim, W.-T. Park, Y.-Y. Noh, J.-J. Kim, Adv. Funct. Mater. 2015, 25, 758.
[75] T. Kurosawa, Y.-C. Chiu, Y. Zhou, X. Gu, W.-C. Chen, Z. Bao, Adv.
Funct. Mater. 2015, 6, 1261.
[76] R. Di Pietro, D. Fazzi, T. B. Kehoe, H. Sirringhaus, J. Am. Chem. Soc.
2012, 134, 14877.
[77] R. Coehoorn, P. A. Bobbert, Phys. Status Solidi A 2012, 209, 2354. [78] C. J. Takacs, N. D. Treat, S. Krämer, Z. Chen, A. Facchetti,
M. L. Chabinyc, A. J. Heeger, Nano Lett. 2013, 13, 2522.
[79] X. Zhang, H. Bronstein, A. J. Kronemeijer, J. Smith, Y. Kim, R. J. Kline, L. J. Richter, T. D. Anthopoulos, H. Sirringhaus, K. Song, M. Heeney, W. Zhang, I. McCulloch, D. M. DeLongchamp, Nat.
Commun. 2013, 4, 2238.
[80] A. Luzio, L. Criante, V. D’Innocenzo, M. Caironi, Sci. Rep. 2013, 3, 3425.
[81] S. M. Sze, K. K. Ng, Physics of Semiconductor Devices, 3rd ed., John Wiley & Sons, Inc., Hoboken, NJ, USA 2006.
[82] J. Ilavsky, J. Appl. Crystallogr. 2012, 45, 324.
[83] B. Watts, L. Thomsen, P. C. Dastoor, J. Electron. Spectrosc. Relat.
Phenom. 2006, 151, 105.
[84] E. Gann, C. R. McNeill, A. Tadich, B. C. C. Cowie, L. Thomsen,
J. Synchrotron Radiat. 2016, 23, 374.
[85] N. M. Kirby, S. T. Mudie, A. M. Hawley, D. J. Cookson, H. D. T. Mertens, N. Cowieson, V. Samardzic-Boban, J. Appl.
Crystal-logr. 2013, 46, 1670.
[86] B. C. C. Cowie, A. Tadich, L. Thomsen, AIP Conf. Proc. 2010, 1234, 307.
